[HDFS-4186] logSync() is called with the write lock held while releasing lease

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 0.23.4, 2.0.2-alpha
• Fix Version/s: 2.0.3-alpha, 0.23.5, 0.23.6
• Component/s: namenode
• Target Version/s: 2.0.3-alpha, 0.23.6
• Hadoop Flags: Reviewed
• Assignee: Kihwal Lee
• Reporter: Kihwal Lee
• Created: 14/Nov/12 02:31
• Updated: 03/Feb/16 03:27
• Resolved: 15/Nov/12 20:46

Description

As pointed out in HDFS-4183, when the lease monitor calls internalReleaseLease(), it acquires the namespace write lock. Inside internalReleaseLease(), if a block recovery is needed, the lease is reassigned to the namenode itself and this is logged & synced in logReassignLease().

Since this is done while the write lock is held, log syncing is blocked. When a large number of leases are expired and blocks are recovered, namenode can slow down.

Issue Links

• relates to: HDFS-4777 File creation with overwrite flag set to true results in logSync holding namesystem lock
• relates to: HDFS-4479 logSync() with the FSNamesystem lock held in commitBlockSynchronization